Advanced Business Analytics using R Project
Mihir Raikar
4/16/2019
library(data.table)## Warning: package 'data.table' was built under R version 3.5.2library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:data.table':
##
## between, first, last## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(tidyverse)## -- Attaching packages ---------------------------------- tidyverse 1.2.1 --## v ggplot2 3.0.0 v readr 1.1.1
## v tibble 1.4.2 v purrr 0.2.5
## v tidyr 0.8.1 v stringr 1.3.1
## v ggplot2 3.0.0 v forcats 0.3.0## -- Conflicts ------------------------------------- tidyverse_conflicts() --
## x dplyr::between() masks data.table::between()
## x dplyr::filter() masks stats::filter()
## x dplyr::first() masks data.table::first()
## x dplyr::lag() masks stats::lag()
## x dplyr::last() masks data.table::last()
## x purrr::transpose() masks data.table::transpose()library(ggplot2)
library(e1071)## Warning: package 'e1071' was built under R version 3.5.2library(caret)## Loading required package: lattice##
## Attaching package: 'caret'## The following object is masked from 'package:purrr':
##
## liftlibrary(tidyverse)
library(gains)
library(corrplot)## corrplot 0.84 loadedlibrary(mefa4)## Loading required package: Matrix##
## Attaching package: 'Matrix'## The following object is masked from 'package:tidyr':
##
## expand## Loading required package: pbapply## Warning: package 'pbapply' was built under R version 3.5.2## mefa4 0.3-5 2018-03-24library(randomForest)## randomForest 4.6-14## Type rfNews() to see new features/changes/bug fixes.##
## Attaching package: 'randomForest'## The following object is masked from 'package:ggplot2':
##
## margin## The following object is masked from 'package:dplyr':
##
## combinelibrary(glmnet)## Loading required package: foreach##
## Attaching package: 'foreach'## The following objects are masked from 'package:purrr':
##
## accumulate, when## Loaded glmnet 2.0-16library(corrplot)
library(gridExtra)##
## Attaching package: 'gridExtra'## The following object is masked from 'package:randomForest':
##
## combine## The following object is masked from 'package:dplyr':
##
## combinelibrary(arm)## Loading required package: MASS##
## Attaching package: 'MASS'## The following object is masked from 'package:dplyr':
##
## select## Loading required package: lme4##
## arm (Version 1.10-1, built: 2018-4-12)## Working directory is /Users/mihirraikar/Downloads##
## Attaching package: 'arm'## The following object is masked from 'package:corrplot':
##
## corrplotlibrary(GGally)##
## Attaching package: 'GGally'## The following object is masked from 'package:dplyr':
##
## nasalibrary(car)## Loading required package: carData##
## Attaching package: 'car'## The following object is masked from 'package:arm':
##
## logit## The following object is masked from 'package:purrr':
##
## some## The following object is masked from 'package:dplyr':
##
## recodelibrary(ggcorrplot)wine <- fread("winequality-white.csv")sum(duplicated(wine))## [1] 0dim(wine)## [1] 4898 13summary(wine)## Obs fixed_acidity volatile_acidity citric_acid
## Min. : 1 Min. : 3.800 Min. :0.0800 Min. :0.0000
## 1st Qu.:1225 1st Qu.: 6.300 1st Qu.:0.2100 1st Qu.:0.2700
## Median :2450 Median : 6.800 Median :0.2600 Median :0.3200
## Mean :2450 Mean : 6.855 Mean :0.2782 Mean :0.3342
## 3rd Qu.:3674 3rd Qu.: 7.300 3rd Qu.:0.3200 3rd Qu.:0.3900
## Max. :4898 Max. :14.200 Max. :1.1000 Max. :1.6600
## residual_sugar chlorides free_sulfur_dioxide
## Min. : 0.600 Min. :0.00900 Min. : 2.00
## 1st Qu.: 1.700 1st Qu.:0.03600 1st Qu.: 23.00
## Median : 5.200 Median :0.04300 Median : 34.00
## Mean : 6.391 Mean :0.04577 Mean : 35.31
## 3rd Qu.: 9.900 3rd Qu.:0.05000 3rd Qu.: 46.00
## Max. :65.800 Max. :0.34600 Max. :289.00
## total_sulfur_dioxide density pH sulphates
## Min. : 9.0 Min. :0.9871 Min. :2.720 Min. :0.2200
## 1st Qu.:108.0 1st Qu.:0.9917 1st Qu.:3.090 1st Qu.:0.4100
## Median :134.0 Median :0.9937 Median :3.180 Median :0.4700
## Mean :138.4 Mean :0.9940 Mean :3.188 Mean :0.4898
## 3rd Qu.:167.0 3rd Qu.:0.9961 3rd Qu.:3.280 3rd Qu.:0.5500
## Max. :440.0 Max. :1.0390 Max. :3.820 Max. :1.0800
## alcohol quality
## Min. : 8.00 Min. :3.000
## 1st Qu.: 9.50 1st Qu.:5.000
## Median :10.40 Median :6.000
## Mean :10.51 Mean :5.878
## 3rd Qu.:11.40 3rd Qu.:6.000
## Max. :14.20 Max. :9.000Our Dependent Variable is Quality, hence we will take a look at the distribution of quality
quality_plot <- ggplot(aes(quality), data = wine) + geom_bar() + ggtitle ("Quality chart")
quality_plot
The range of Quality is from 3 to 9. We can bucket the data into good, medium and bad with reference to Quality
wine$quality_level <- ifelse(wine$quality >= 7, "good", NA)
wine$quality_level <- ifelse(wine$quality <= 6, "medium", wine$quality_level)
wine$quality_level <- ifelse(wine$quality <= 4, "bad", wine$quality_level)
wine$quality_level <- as.factor(wine$quality_level)
wine$quality <- as.factor(wine$quality)Plot and Visualize Quality level
qualitylevel_plot <- ggplot(aes(quality,fill=quality_level),data=wine) + geom_bar() + ggtitle ("Quality level chart")
qualitylevel_plot
qualitylevelcount_plot<-qplot(wine$quality_level) + xlab("quality level") + ggtitle("count of quality level")
grid.arrange(quality_plot,qualitylevel_plot,ncol=2)
We will now explore various the variables
summary(wine$fixed_acidity)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3.800 6.300 6.800 6.855 7.300 14.200ggplot(aes(fixed_acidity), data = wine) + geom_bar()
ggplot(aes(x=quality, y=fixed_acidity), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=fixed_acidity), data = wine) + geom_boxplot()
The data is positively skewed and also has some outliers The average fixed acidity tends to decrease with increase in quality but increases with quality=9 The observations of fixed acidity at it has in general a negative relationship with quality level
summary(wine$volatile_acidity)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0800 0.2100 0.2600 0.2782 0.3200 1.1000ggplot(aes(volatile_acidity), data = wine) + geom_bar()
ggplot(aes(x=quality, y=volatile_acidity), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=volatile_acidity), data = wine) + geom_boxplot()
Similar to fixed acidity, volatile acidity is also positive skewed with outliers but the range is very small The relationship with quality is unclear as there is no trend The relationship of volatile acidity with quality level is negative
summary(wine$citric_acid)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0000 0.2700 0.3200 0.3342 0.3900 1.6600ggplot(aes(citric_acid), data = wine) + geom_bar()
ggplot(aes(x=quality, y=citric_acid), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=citric_acid), data = wine) + geom_boxplot()
The distribution of citric acid is similar to normal but has 2 unusual peaks and a few outliers. The peaks are causing the data to not have a general trend
summary(wine$residual_sugar)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.600 1.700 5.200 6.391 9.900 65.800ggplot(aes(residual_sugar), data = wine) + geom_bar()
ggplot(aes(x=quality, y=residual_sugar), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=residual_sugar), data = wine) + geom_boxplot()
The data is positively skewed. It looks like most of the wines have very low residual sugars It looks like residual sugar is low in bad wines but comaritively high in medium and again comparatively lower in good. So generally, residual sugar in wine is good, but really good wines dont have it as much as average wines
summary(wine$chlorides)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00900 0.03600 0.04300 0.04577 0.05000 0.34600ggplot(aes(chlorides), data = wine) + geom_bar()
ggplot(aes(x=quality, y=chlorides), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=chlorides), data = wine) + geom_boxplot()
Chlorides are also positively skewed. The max value is lot higher than mean. Chlorides also have a negative relationship with quality level but not very significant
summary(wine$free_sulfur_dioxide)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 2.00 23.00 34.00 35.31 46.00 289.00ggplot(aes(free_sulfur_dioxide), data = wine) + geom_bar()
ggplot(aes(x=quality, y=free_sulfur_dioxide), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=free_sulfur_dioxide), data = wine) + geom_boxplot()
There is a huge outlier(289) whereas the rest of the data is below 150. Apart from the outlier also, the data is a little positive skewed. There is a positive relationship with quality but is not that significant with good quality
summary(wine$total_sulfur_dioxide)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 9.0 108.0 134.0 138.4 167.0 440.0ggplot(aes(total_sulfur_dioxide), data = wine) + geom_bar()
ggplot(aes(x=quality, y=total_sulfur_dioxide), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=total_sulfur_dioxide), data = wine) + geom_boxplot()
There are some peaks but the distribution is similar to normal Similar to residual sugar, the bad wines have low total sulphur dioxide, medium has high but good have lower than medium
summary(wine$density)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.9871 0.9917 0.9937 0.9940 0.9961 1.0390ggplot(aes(density), data = wine) + geom_bar()
ggplot(aes(x=quality, y=density), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=density), data = wine) + geom_boxplot()
The range is very low and has a couple of outliers There is a negative relationship between density and quality level
summary(wine$volatile_acidity)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0800 0.2100 0.2600 0.2782 0.3200 1.1000ggplot(aes(pH), data = wine) + geom_bar()
ggplot(aes(x=quality, y=pH), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=pH), data = wine) + geom_boxplot()
pH has a distribution very close to normal. there are peaks but no significant outliers There is a clear positive relationship between pH and Quality/Quality level.
summary(wine$sulphates)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.2200 0.4100 0.4700 0.4898 0.5500 1.0800ggplot(aes(sulphates), data = wine) + geom_bar()
ggplot(aes(x=quality, y=sulphates), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=sulphates), data = wine) + geom_boxplot()
The data is positive skewed and a few high values Clear positive relation with quality
summary(wine$alcohol)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 8.00 9.50 10.40 10.51 11.40 14.20ggplot(aes(alcohol), data = wine) + geom_bar()
ggplot(aes(x=quality, y=alcohol), data = wine) + geom_boxplot()
ggplot(aes(x=quality_level, y=alcohol), data = wine) + geom_boxplot()
There are a number of peaks, looks like no outliers There seems to be a positive relationship with quality butthe trend is not completely clear. Good wines have very high alchohol but medium wines have lower alchohol than bad wines which is surprising
ggcorr(wine)## Warning in ggcorr(wine): data in column(s) 'quality', 'quality_level' are
## not numeric and were ignored
Density and Residual sugar has really high correlation (0.84) Density and alchohol also has a high (negative) correlation (-0.78) Density also as a high correlation with total sulfur dioxide(0.53) Free sulfur dioxide and total sulfur dioxide are also highly correlated(0.62) Alchohol is negatively correlated with total sulfur dioxide and residual sugar(both -45) Alchohol and quality has a correlation of 44 pH and fixed acidity are also negatively correlated (-0.43)
set.seed(123)
samp <- sample(nrow(wine), 0.6 * nrow(wine))
train <- wine[samp, ]
test <- wine[-samp, ]svmlinear1 <- train(quality ~., data = train,
method = "svmLinear",
trControl=trainControl(method = "repeatedcv",
number = 10, repeats = 10),
preProcess = c("center", "scale"),
tuneLength = 10)
svmlinear1$bestTune## C
## 1 1summary(svmlinear1)## Length Class Mode
## 1 ksvm S4predsvmLinear1 <- predict(svmlinear1, test)
confusionMatrix(table(predsvmLinear1, test$quality))## Confusion Matrix and Statistics
##
##
## predsvmLinear1 3 4 5 6 7 8 9
## 3 0 1 0 0 0 0 0
## 4 6 72 0 0 0 0 0
## 5 0 0 304 139 0 0 0
## 6 0 0 284 743 0 0 0
## 7 0 0 0 0 337 73 1
## 8 0 0 0 0 0 0 0
## 9 0 0 0 0 0 0 0
##
## Overall Statistics
##
## Accuracy : 0.7429
## 95% CI : (0.7229, 0.7621)
## No Information Rate : 0.45
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.6097
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: 3 Class: 4 Class: 5 Class: 6 Class: 7
## Sensitivity 0.0000000 0.98630 0.5170 0.8424 1.0000
## Specificity 0.9994882 0.99682 0.8987 0.7365 0.9544
## Pos Pred Value 0.0000000 0.92308 0.6862 0.7235 0.8200
## Neg Pred Value 0.9969372 0.99947 0.8128 0.8510 1.0000
## Prevalence 0.0030612 0.03724 0.3000 0.4500 0.1719
## Detection Rate 0.0000000 0.03673 0.1551 0.3791 0.1719
## Detection Prevalence 0.0005102 0.03980 0.2260 0.5240 0.2097
## Balanced Accuracy 0.4997441 0.99156 0.7078 0.7895 0.9772
## Class: 8 Class: 9
## Sensitivity 0.00000 0.0000000
## Specificity 1.00000 1.0000000
## Pos Pred Value NaN NaN
## Neg Pred Value 0.96276 0.9994898
## Prevalence 0.03724 0.0005102
## Detection Rate 0.00000 0.0000000
## Detection Prevalence 0.00000 0.0000000
## Balanced Accuracy 0.50000 0.5000000We have run the SVM model on Quality variable. For the best model, we found C=1. The accuracy is 74.29%
svmlinear2 <- train(quality_level ~., data = train,
method = "svmLinear",
trControl=trainControl(method = "repeatedcv",
number = 10, repeats = 10),
preProcess = c("center", "scale"),
tuneLength = 10)
svmlinear2$bestTune## C
## 1 1summary(svmlinear2)## Length Class Mode
## 1 ksvm S4predsvmLinear2 <- predict(svmlinear2, test)
confusionMatrix(table(predsvmLinear2, test$quality_level))## Confusion Matrix and Statistics
##
##
## predsvmLinear2 bad good medium
## bad 79 0 0
## good 0 411 0
## medium 0 0 1470
##
## Overall Statistics
##
## Accuracy : 1
## 95% CI : (0.9981, 1)
## No Information Rate : 0.75
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 1
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: bad Class: good Class: medium
## Sensitivity 1.00000 1.0000 1.00
## Specificity 1.00000 1.0000 1.00
## Pos Pred Value 1.00000 1.0000 1.00
## Neg Pred Value 1.00000 1.0000 1.00
## Prevalence 0.04031 0.2097 0.75
## Detection Rate 0.04031 0.2097 0.75
## Detection Prevalence 0.04031 0.2097 0.75
## Balanced Accuracy 1.00000 1.0000 1.00We have run the SVM model on Quality Level variable. For the best model, we found C=1. The accuracy is 100%
nb1 <- naiveBayes(quality ~ ., data = train)
nb1##
## Naive Bayes Classifier for Discrete Predictors
##
## Call:
## naiveBayes.default(x = X, y = Y, laplace = laplace)
##
## A-priori probabilities:
## Y
## 3 4 5 6 7 8
## 0.004765146 0.030633084 0.295779442 0.447923758 0.184819605 0.034717495
## 9
## 0.001361470
##
## Conditional probabilities:
## Obs
## Y [,1] [,2]
## 3 2099.929 1376.8218
## 4 2225.933 1358.8583
## 5 2348.731 1389.2656
## 6 2572.020 1443.3965
## 7 2435.808 1420.9248
## 8 2308.284 1301.7364
## 9 1021.500 391.8865
##
## fixed_acidity
## Y [,1] [,2]
## 3 7.085714 1.3283055
## 4 7.224444 1.1396147
## 5 6.952589 0.8623638
## 6 6.853305 0.8474624
## 7 6.744015 0.7628095
## 8 6.683333 0.7880397
## 9 7.625000 1.0045729
##
## volatile_acidity
## Y [,1] [,2]
## 3 0.3425000 0.14183888
## 4 0.3950000 0.17425604
## 5 0.3030667 0.10381436
## 6 0.2630471 0.09024842
## 7 0.2637385 0.09060641
## 8 0.2681863 0.10808135
## 9 0.2825000 0.05315073
##
## citric_acid
## Y [,1] [,2]
## 3 0.3142857 0.08327104
## 4 0.3047778 0.16604607
## 5 0.3403452 0.14220570
## 6 0.3371657 0.12126998
## 7 0.3270350 0.07921812
## 8 0.3280392 0.07226244
## 9 0.4100000 0.07164728
##
## residual_sugar
## Y [,1] [,2]
## 3 7.207143 5.383338
## 4 4.403889 3.824766
## 5 7.318297 5.359412
## 6 6.571429 5.264301
## 7 5.126611 4.296436
## 8 5.754902 4.428405
## 9 4.750000 4.024508
##
## chlorides
## Y [,1] [,2]
## 3 0.05907143 0.054862716
## 4 0.05053333 0.030955278
## 5 0.05108516 0.025828611
## 6 0.04520289 0.020006136
## 7 0.03804788 0.010193040
## 8 0.03855882 0.012628409
## 9 0.02900000 0.007527727
##
## free_sulfur_dioxide
## Y [,1] [,2]
## 3 68.17857 78.83688
## 4 21.83889 17.50246
## 5 36.11795 17.88182
## 6 35.50684 15.31668
## 7 34.41344 13.58294
## 8 37.03922 17.56793
## 9 35.75000 14.26826
##
## total_sulfur_dioxide
## Y [,1] [,2]
## 3 200.0714 113.36715
## 4 122.9444 50.84088
## 5 149.4689 43.69120
## 6 137.5794 41.25911
## 7 124.9429 33.02970
## 8 126.9314 33.33679
## 9 123.7500 11.11680
##
## density
## Y [,1] [,2]
## 3 0.9950629 0.002737872
## 4 0.9942519 0.002343708
## 5 0.9952456 0.002533501
## 6 0.9940534 0.003078789
## 7 0.9924701 0.002811543
## 8 0.9924160 0.002822665
## 9 0.9919125 0.003405969
##
## pH
## Y [,1] [,2]
## 3 3.206429 0.22165834
## 4 3.164778 0.16100267
## 5 3.166605 0.13844631
## 6 3.187302 0.15220639
## 7 3.211142 0.15931757
## 8 3.213431 0.14482830
## 9 3.282500 0.06946222
##
## sulphates
## Y [,1] [,2]
## 3 0.5092857 0.11505613
## 4 0.4713333 0.13418250
## 5 0.4817722 0.09809165
## 6 0.4905243 0.11392860
## 7 0.5075322 0.13382875
## 8 0.4920588 0.14982061
## 9 0.4300000 0.05291503
##
## alcohol
## Y [,1] [,2]
## 3 10.300000 1.179309
## 4 10.142222 1.015495
## 5 9.812712 0.845418
## 6 10.552685 1.122599
## 7 11.336108 1.263374
## 8 11.511765 1.284584
## 9 12.125000 1.161536
##
## quality_level
## Y bad good medium
## 3 1 0 0
## 4 1 0 0
## 5 0 0 1
## 6 0 0 1
## 7 0 1 0
## 8 0 1 0
## 9 0 1 0# probabilities
pred.prob <- predict(nb1, newdata = test, type = "raw")
# class membership
pred.class1 <- predict(nb1, newdata = test)
confusionMatrix(pred.class1, test$quality)## Confusion Matrix and Statistics
##
## Reference
## Prediction 3 4 5 6 7 8 9
## 3 0 2 4 1 0 0 0
## 4 6 64 2 1 0 0 0
## 5 0 7 341 234 10 3 0
## 6 0 0 241 646 0 1 0
## 7 0 0 0 0 324 67 1
## 8 0 0 0 0 3 1 0
## 9 0 0 0 0 0 1 0
##
## Overall Statistics
##
## Accuracy : 0.702
## 95% CI : (0.6812, 0.7222)
## No Information Rate : 0.45
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.5547
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: 3 Class: 4 Class: 5 Class: 6 Class: 7
## Sensitivity 0.000000 0.87671 0.5799 0.7324 0.9614
## Specificity 0.996418 0.99523 0.8149 0.7755 0.9581
## Pos Pred Value 0.000000 0.87671 0.5731 0.7275 0.8265
## Neg Pred Value 0.996928 0.99523 0.8190 0.7799 0.9917
## Prevalence 0.003061 0.03724 0.3000 0.4500 0.1719
## Detection Rate 0.000000 0.03265 0.1740 0.3296 0.1653
## Detection Prevalence 0.003571 0.03724 0.3036 0.4531 0.2000
## Balanced Accuracy 0.498209 0.93597 0.6974 0.7540 0.9598
## Class: 8 Class: 9
## Sensitivity 0.0136986 0.0000000
## Specificity 0.9984102 0.9994895
## Pos Pred Value 0.2500000 0.0000000
## Neg Pred Value 0.9631902 0.9994895
## Prevalence 0.0372449 0.0005102
## Detection Rate 0.0005102 0.0000000
## Detection Prevalence 0.0020408 0.0005102
## Balanced Accuracy 0.5060544 0.4997448We have run the Naive Bayes model on Quality variable. The accuracy is 70.2%
nb2 <- naiveBayes(quality_level ~ ., data = train)
nb2##
## Naive Bayes Classifier for Discrete Predictors
##
## Call:
## naiveBayes.default(x = X, y = Y, laplace = laplace)
##
## A-priori probabilities:
## Y
## bad good medium
## 0.03539823 0.22089857 0.74370320
##
## Conditional probabilities:
## Obs
## Y [,1] [,2]
## bad 2208.971 1355.226
## good 2407.049 1402.733
## medium 2483.215 1425.990
##
## fixed_acidity
## Y [,1] [,2]
## bad 7.205769 1.1606683
## good 6.739908 0.7703979
## medium 6.892792 0.8546064
##
## volatile_acidity
## Y [,1] [,2]
## bad 0.3879327 0.17059094
## good 0.2645532 0.09330084
## medium 0.2789634 0.09783234
##
## citric_acid
## Y [,1] [,2]
## bad 0.3060577 0.15719272
## good 0.3277042 0.07828682
## medium 0.3384302 0.12997836
##
## residual_sugar
## Y [,1] [,2]
## bad 4.781250 4.149998
## good 5.223035 4.315678
## medium 6.868467 5.313702
##
## chlorides
## Y [,1] [,2]
## bad 0.05168269 0.03487767
## good 0.03807242 0.01060977
## medium 0.04754233 0.02268080
##
## free_sulfur_dioxide
## Y [,1] [,2]
## bad 28.07692 36.07952
## good 34.83436 14.29274
## medium 35.74989 16.38377
##
## total_sulfur_dioxide
## Y [,1] [,2]
## bad 133.3269 67.49278
## good 125.2481 32.96710
## medium 142.3080 42.63243
##
## density
## Y [,1] [,2]
## bad 0.9943611 0.002402041
## good 0.9924581 0.002812382
## medium 0.9945276 0.002932400
##
## pH
## Y [,1] [,2]
## bad 3.170385 0.1697166
## good 3.211941 0.1566950
## medium 3.179071 0.1472051
##
## sulphates
## Y [,1] [,2]
## bad 0.4764423 0.1319011
## good 0.5046225 0.1362286
## medium 0.4870435 0.1079706
##
## alcohol
## Y [,1] [,2]
## bad 10.16346 1.034179
## good 11.36858 1.267324
## medium 10.25839 1.083546
##
## quality
## Y 3 4 5 6 7
## bad 0.134615385 0.865384615 0.000000000 0.000000000 0.000000000
## good 0.000000000 0.000000000 0.000000000 0.000000000 0.836671803
## medium 0.000000000 0.000000000 0.397711670 0.602288330 0.000000000
## quality
## Y 8 9
## bad 0.000000000 0.000000000
## good 0.157164869 0.006163328
## medium 0.000000000 0.000000000# probabilities
pred.prob <- predict(nb2, newdata = test, type = "raw")
# class membership
pred.class2 <- predict(nb2, newdata = test)
confusionMatrix(pred.class2, test$quality_level)## Confusion Matrix and Statistics
##
## Reference
## Prediction bad good medium
## bad 72 1 16
## good 2 393 2
## medium 5 17 1452
##
## Overall Statistics
##
## Accuracy : 0.9781
## 95% CI : (0.9706, 0.9841)
## No Information Rate : 0.75
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.944
## Mcnemar's Test P-Value : 0.0004531
##
## Statistics by Class:
##
## Class: bad Class: good Class: medium
## Sensitivity 0.91139 0.9562 0.9878
## Specificity 0.99096 0.9974 0.9551
## Pos Pred Value 0.80899 0.9899 0.9851
## Neg Pred Value 0.99626 0.9885 0.9630
## Prevalence 0.04031 0.2097 0.7500
## Detection Rate 0.03673 0.2005 0.7408
## Detection Prevalence 0.04541 0.2026 0.7520
## Balanced Accuracy 0.95118 0.9768 0.9714We have run the Naive Bayes model on Quality Level variable. The accuracy is 97.81%
set.seed(123)
ridgeReg1 <- train(quality~., train, method = 'glmnet',
tuneGrid = expand.grid(alpha = 0,
lambda = seq(0.0001, 0.5, length = 5)),
trControl = trainControl(method = "repeatedcv",
number = 10, repeats = 10,
verboseIter = TRUE))## + Fold01.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep01: alpha=0, lambda=0.5
## + Fold02.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep01: alpha=0, lambda=0.5
## + Fold03.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep01: alpha=0, lambda=0.5
## + Fold04.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep01: alpha=0, lambda=0.5
## + Fold05.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep01: alpha=0, lambda=0.5
## + Fold06.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep01: alpha=0, lambda=0.5
## + Fold07.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep01: alpha=0, lambda=0.5
## + Fold08.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep01: alpha=0, lambda=0.5
## + Fold09.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep01: alpha=0, lambda=0.5
## + Fold10.Rep01: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep01: alpha=0, lambda=0.5
## + Fold01.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep02: alpha=0, lambda=0.5
## + Fold02.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep02: alpha=0, lambda=0.5
## + Fold03.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep02: alpha=0, lambda=0.5
## + Fold04.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep02: alpha=0, lambda=0.5
## + Fold05.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep02: alpha=0, lambda=0.5
## + Fold06.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep02: alpha=0, lambda=0.5
## + Fold07.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep02: alpha=0, lambda=0.5
## + Fold08.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep02: alpha=0, lambda=0.5
## + Fold09.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep02: alpha=0, lambda=0.5
## + Fold10.Rep02: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep02: alpha=0, lambda=0.5
## + Fold01.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep03: alpha=0, lambda=0.5
## + Fold02.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep03: alpha=0, lambda=0.5
## + Fold03.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep03: alpha=0, lambda=0.5
## + Fold04.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep03: alpha=0, lambda=0.5
## + Fold05.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep03: alpha=0, lambda=0.5
## + Fold06.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep03: alpha=0, lambda=0.5
## + Fold07.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep03: alpha=0, lambda=0.5
## + Fold08.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep03: alpha=0, lambda=0.5
## + Fold09.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep03: alpha=0, lambda=0.5
## + Fold10.Rep03: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep03: alpha=0, lambda=0.5
## + Fold01.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep04: alpha=0, lambda=0.5
## + Fold02.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep04: alpha=0, lambda=0.5
## + Fold03.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep04: alpha=0, lambda=0.5
## + Fold04.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep04: alpha=0, lambda=0.5
## + Fold05.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep04: alpha=0, lambda=0.5
## + Fold06.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep04: alpha=0, lambda=0.5
## + Fold07.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep04: alpha=0, lambda=0.5
## + Fold08.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep04: alpha=0, lambda=0.5
## + Fold09.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep04: alpha=0, lambda=0.5
## + Fold10.Rep04: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep04: alpha=0, lambda=0.5
## + Fold01.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep05: alpha=0, lambda=0.5
## + Fold02.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep05: alpha=0, lambda=0.5
## + Fold03.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep05: alpha=0, lambda=0.5
## + Fold04.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep05: alpha=0, lambda=0.5
## + Fold05.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep05: alpha=0, lambda=0.5
## + Fold06.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep05: alpha=0, lambda=0.5
## + Fold07.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep05: alpha=0, lambda=0.5
## + Fold08.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep05: alpha=0, lambda=0.5
## + Fold09.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep05: alpha=0, lambda=0.5
## + Fold10.Rep05: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep05: alpha=0, lambda=0.5
## + Fold01.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep06: alpha=0, lambda=0.5
## + Fold02.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep06: alpha=0, lambda=0.5
## + Fold03.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep06: alpha=0, lambda=0.5
## + Fold04.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep06: alpha=0, lambda=0.5
## + Fold05.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep06: alpha=0, lambda=0.5
## + Fold06.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep06: alpha=0, lambda=0.5
## + Fold07.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep06: alpha=0, lambda=0.5
## + Fold08.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep06: alpha=0, lambda=0.5
## + Fold09.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep06: alpha=0, lambda=0.5
## + Fold10.Rep06: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep06: alpha=0, lambda=0.5
## + Fold01.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep07: alpha=0, lambda=0.5
## + Fold02.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep07: alpha=0, lambda=0.5
## + Fold03.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep07: alpha=0, lambda=0.5
## + Fold04.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep07: alpha=0, lambda=0.5
## + Fold05.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep07: alpha=0, lambda=0.5
## + Fold06.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep07: alpha=0, lambda=0.5
## + Fold07.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep07: alpha=0, lambda=0.5
## + Fold08.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep07: alpha=0, lambda=0.5
## + Fold09.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep07: alpha=0, lambda=0.5
## + Fold10.Rep07: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep07: alpha=0, lambda=0.5
## + Fold01.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep08: alpha=0, lambda=0.5
## + Fold02.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep08: alpha=0, lambda=0.5
## + Fold03.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep08: alpha=0, lambda=0.5
## + Fold04.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep08: alpha=0, lambda=0.5
## + Fold05.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep08: alpha=0, lambda=0.5
## + Fold06.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep08: alpha=0, lambda=0.5
## + Fold07.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep08: alpha=0, lambda=0.5
## + Fold08.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep08: alpha=0, lambda=0.5
## + Fold09.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep08: alpha=0, lambda=0.5
## + Fold10.Rep08: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep08: alpha=0, lambda=0.5
## + Fold01.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep09: alpha=0, lambda=0.5
## + Fold02.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep09: alpha=0, lambda=0.5
## + Fold03.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep09: alpha=0, lambda=0.5
## + Fold04.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep09: alpha=0, lambda=0.5
## + Fold05.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep09: alpha=0, lambda=0.5
## + Fold06.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep09: alpha=0, lambda=0.5
## + Fold07.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep09: alpha=0, lambda=0.5
## + Fold08.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep09: alpha=0, lambda=0.5
## + Fold09.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep09: alpha=0, lambda=0.5
## + Fold10.Rep09: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep09: alpha=0, lambda=0.5
## + Fold01.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep10: alpha=0, lambda=0.5
## + Fold02.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep10: alpha=0, lambda=0.5
## + Fold03.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep10: alpha=0, lambda=0.5
## + Fold04.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep10: alpha=0, lambda=0.5
## + Fold05.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep10: alpha=0, lambda=0.5
## + Fold06.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep10: alpha=0, lambda=0.5
## + Fold07.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep10: alpha=0, lambda=0.5
## + Fold08.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep10: alpha=0, lambda=0.5
## + Fold09.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep10: alpha=0, lambda=0.5
## + Fold10.Rep10: alpha=0, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep10: alpha=0, lambda=0.5
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0, lambda = 1e-04 on full training set## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground# print results
print(ridgeReg1)## glmnet
##
## 2938 samples
## 13 predictor
## 7 classes: '3', '4', '5', '6', '7', '8', '9'
##
## No pre-processing
## Resampling: Cross-Validated (10 fold, repeated 10 times)
## Summary of sample sizes: 2645, 2645, 2645, 2644, 2644, 2644, ...
## Resampling results across tuning parameters:
##
## lambda Accuracy Kappa
## 0.000100 0.7151884 0.5642168
## 0.125075 0.6875257 0.5115899
## 0.250050 0.6754056 0.4874629
## 0.375025 0.6628432 0.4629472
## 0.500000 0.6589265 0.4533713
##
## Tuning parameter 'alpha' was held constant at a value of 0
## Accuracy was used to select the optimal model using the largest value.
## The final values used for the model were alpha = 0 and lambda = 1e-04.# plot results
plot(ridgeReg1)
plot(varImp(ridgeReg1, scale = FALSE))
plot(varImp(ridgeReg1, scale = TRUE))
PredictRidge1 <- predict(ridgeReg1, test)
confusionMatrix(PredictRidge1, test$quality)## Confusion Matrix and Statistics
##
## Reference
## Prediction 3 4 5 6 7 8 9
## 3 0 0 0 0 0 0 0
## 4 4 46 0 0 0 0 0
## 5 0 15 278 132 0 0 0
## 6 2 12 310 750 0 0 0
## 7 0 0 0 0 337 73 1
## 8 0 0 0 0 0 0 0
## 9 0 0 0 0 0 0 0
##
## Overall Statistics
##
## Accuracy : 0.7199
## 95% CI : (0.6994, 0.7397)
## No Information Rate : 0.45
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.57
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: 3 Class: 4 Class: 5 Class: 6 Class: 7 Class: 8
## Sensitivity 0.000000 0.63014 0.4728 0.8503 1.0000 0.00000
## Specificity 1.000000 0.99788 0.8929 0.6994 0.9544 1.00000
## Pos Pred Value NaN 0.92000 0.6541 0.6983 0.8200 NaN
## Neg Pred Value 0.996939 0.98586 0.7980 0.8510 1.0000 0.96276
## Prevalence 0.003061 0.03724 0.3000 0.4500 0.1719 0.03724
## Detection Rate 0.000000 0.02347 0.1418 0.3827 0.1719 0.00000
## Detection Prevalence 0.000000 0.02551 0.2168 0.5480 0.2097 0.00000
## Balanced Accuracy 0.500000 0.81401 0.6828 0.7749 0.9772 0.50000
## Class: 9
## Sensitivity 0.0000000
## Specificity 1.0000000
## Pos Pred Value NaN
## Neg Pred Value 0.9994898
## Prevalence 0.0005102
## Detection Rate 0.0000000
## Detection Prevalence 0.0000000
## Balanced Accuracy 0.5000000We have run the Ridge model on Quality variable. The accuracy is 71.99%
set.seed(123)
ridgeReg2 <- train(quality_level~., train, method = 'glmnet',
tuneGrid = expand.grid(alpha = 0,
lambda = seq(0.0001, 0.5, length = 5)),
trControl = trainControl(method = "repeatedcv",
number = 10, repeats = 10,
verboseIter = TRUE))## + Fold01.Rep01: alpha=0, lambda=0.5
## - Fold01.Rep01: alpha=0, lambda=0.5
## + Fold02.Rep01: alpha=0, lambda=0.5
## - Fold02.Rep01: alpha=0, lambda=0.5
## + Fold03.Rep01: alpha=0, lambda=0.5
## - Fold03.Rep01: alpha=0, lambda=0.5
## + Fold04.Rep01: alpha=0, lambda=0.5
## - Fold04.Rep01: alpha=0, lambda=0.5
## + Fold05.Rep01: alpha=0, lambda=0.5
## - Fold05.Rep01: alpha=0, lambda=0.5
## + Fold06.Rep01: alpha=0, lambda=0.5
## - Fold06.Rep01: alpha=0, lambda=0.5
## + Fold07.Rep01: alpha=0, lambda=0.5
## - Fold07.Rep01: alpha=0, lambda=0.5
## + Fold08.Rep01: alpha=0, lambda=0.5
## - Fold08.Rep01: alpha=0, lambda=0.5
## + Fold09.Rep01: alpha=0, lambda=0.5
## - Fold09.Rep01: alpha=0, lambda=0.5
## + Fold10.Rep01: alpha=0, lambda=0.5
## - Fold10.Rep01: alpha=0, lambda=0.5
## + Fold01.Rep02: alpha=0, lambda=0.5
## - Fold01.Rep02: alpha=0, lambda=0.5
## + Fold02.Rep02: alpha=0, lambda=0.5
## - Fold02.Rep02: alpha=0, lambda=0.5
## + Fold03.Rep02: alpha=0, lambda=0.5
## - Fold03.Rep02: alpha=0, lambda=0.5
## + Fold04.Rep02: alpha=0, lambda=0.5
## - Fold04.Rep02: alpha=0, lambda=0.5
## + Fold05.Rep02: alpha=0, lambda=0.5
## - Fold05.Rep02: alpha=0, lambda=0.5
## + Fold06.Rep02: alpha=0, lambda=0.5
## - Fold06.Rep02: alpha=0, lambda=0.5
## + Fold07.Rep02: alpha=0, lambda=0.5
## - Fold07.Rep02: alpha=0, lambda=0.5
## + Fold08.Rep02: alpha=0, lambda=0.5
## - Fold08.Rep02: alpha=0, lambda=0.5
## + Fold09.Rep02: alpha=0, lambda=0.5
## - Fold09.Rep02: alpha=0, lambda=0.5
## + Fold10.Rep02: alpha=0, lambda=0.5
## - Fold10.Rep02: alpha=0, lambda=0.5
## + Fold01.Rep03: alpha=0, lambda=0.5
## - Fold01.Rep03: alpha=0, lambda=0.5
## + Fold02.Rep03: alpha=0, lambda=0.5
## - Fold02.Rep03: alpha=0, lambda=0.5
## + Fold03.Rep03: alpha=0, lambda=0.5
## - Fold03.Rep03: alpha=0, lambda=0.5
## + Fold04.Rep03: alpha=0, lambda=0.5
## - Fold04.Rep03: alpha=0, lambda=0.5
## + Fold05.Rep03: alpha=0, lambda=0.5
## - Fold05.Rep03: alpha=0, lambda=0.5
## + Fold06.Rep03: alpha=0, lambda=0.5
## - Fold06.Rep03: alpha=0, lambda=0.5
## + Fold07.Rep03: alpha=0, lambda=0.5
## - Fold07.Rep03: alpha=0, lambda=0.5
## + Fold08.Rep03: alpha=0, lambda=0.5
## - Fold08.Rep03: alpha=0, lambda=0.5
## + Fold09.Rep03: alpha=0, lambda=0.5
## - Fold09.Rep03: alpha=0, lambda=0.5
## + Fold10.Rep03: alpha=0, lambda=0.5
## - Fold10.Rep03: alpha=0, lambda=0.5
## + Fold01.Rep04: alpha=0, lambda=0.5
## - Fold01.Rep04: alpha=0, lambda=0.5
## + Fold02.Rep04: alpha=0, lambda=0.5
## - Fold02.Rep04: alpha=0, lambda=0.5
## + Fold03.Rep04: alpha=0, lambda=0.5
## - Fold03.Rep04: alpha=0, lambda=0.5
## + Fold04.Rep04: alpha=0, lambda=0.5
## - Fold04.Rep04: alpha=0, lambda=0.5
## + Fold05.Rep04: alpha=0, lambda=0.5
## - Fold05.Rep04: alpha=0, lambda=0.5
## + Fold06.Rep04: alpha=0, lambda=0.5
## - Fold06.Rep04: alpha=0, lambda=0.5
## + Fold07.Rep04: alpha=0, lambda=0.5
## - Fold07.Rep04: alpha=0, lambda=0.5
## + Fold08.Rep04: alpha=0, lambda=0.5
## - Fold08.Rep04: alpha=0, lambda=0.5
## + Fold09.Rep04: alpha=0, lambda=0.5
## - Fold09.Rep04: alpha=0, lambda=0.5
## + Fold10.Rep04: alpha=0, lambda=0.5
## - Fold10.Rep04: alpha=0, lambda=0.5
## + Fold01.Rep05: alpha=0, lambda=0.5
## - Fold01.Rep05: alpha=0, lambda=0.5
## + Fold02.Rep05: alpha=0, lambda=0.5
## - Fold02.Rep05: alpha=0, lambda=0.5
## + Fold03.Rep05: alpha=0, lambda=0.5
## - Fold03.Rep05: alpha=0, lambda=0.5
## + Fold04.Rep05: alpha=0, lambda=0.5
## - Fold04.Rep05: alpha=0, lambda=0.5
## + Fold05.Rep05: alpha=0, lambda=0.5
## - Fold05.Rep05: alpha=0, lambda=0.5
## + Fold06.Rep05: alpha=0, lambda=0.5
## - Fold06.Rep05: alpha=0, lambda=0.5
## + Fold07.Rep05: alpha=0, lambda=0.5
## - Fold07.Rep05: alpha=0, lambda=0.5
## + Fold08.Rep05: alpha=0, lambda=0.5
## - Fold08.Rep05: alpha=0, lambda=0.5
## + Fold09.Rep05: alpha=0, lambda=0.5
## - Fold09.Rep05: alpha=0, lambda=0.5
## + Fold10.Rep05: alpha=0, lambda=0.5
## - Fold10.Rep05: alpha=0, lambda=0.5
## + Fold01.Rep06: alpha=0, lambda=0.5
## - Fold01.Rep06: alpha=0, lambda=0.5
## + Fold02.Rep06: alpha=0, lambda=0.5
## - Fold02.Rep06: alpha=0, lambda=0.5
## + Fold03.Rep06: alpha=0, lambda=0.5
## - Fold03.Rep06: alpha=0, lambda=0.5
## + Fold04.Rep06: alpha=0, lambda=0.5
## - Fold04.Rep06: alpha=0, lambda=0.5
## + Fold05.Rep06: alpha=0, lambda=0.5
## - Fold05.Rep06: alpha=0, lambda=0.5
## + Fold06.Rep06: alpha=0, lambda=0.5
## - Fold06.Rep06: alpha=0, lambda=0.5
## + Fold07.Rep06: alpha=0, lambda=0.5
## - Fold07.Rep06: alpha=0, lambda=0.5
## + Fold08.Rep06: alpha=0, lambda=0.5
## - Fold08.Rep06: alpha=0, lambda=0.5
## + Fold09.Rep06: alpha=0, lambda=0.5
## - Fold09.Rep06: alpha=0, lambda=0.5
## + Fold10.Rep06: alpha=0, lambda=0.5
## - Fold10.Rep06: alpha=0, lambda=0.5
## + Fold01.Rep07: alpha=0, lambda=0.5
## - Fold01.Rep07: alpha=0, lambda=0.5
## + Fold02.Rep07: alpha=0, lambda=0.5
## - Fold02.Rep07: alpha=0, lambda=0.5
## + Fold03.Rep07: alpha=0, lambda=0.5
## - Fold03.Rep07: alpha=0, lambda=0.5
## + Fold04.Rep07: alpha=0, lambda=0.5
## - Fold04.Rep07: alpha=0, lambda=0.5
## + Fold05.Rep07: alpha=0, lambda=0.5
## - Fold05.Rep07: alpha=0, lambda=0.5
## + Fold06.Rep07: alpha=0, lambda=0.5
## - Fold06.Rep07: alpha=0, lambda=0.5
## + Fold07.Rep07: alpha=0, lambda=0.5
## - Fold07.Rep07: alpha=0, lambda=0.5
## + Fold08.Rep07: alpha=0, lambda=0.5
## - Fold08.Rep07: alpha=0, lambda=0.5
## + Fold09.Rep07: alpha=0, lambda=0.5
## - Fold09.Rep07: alpha=0, lambda=0.5
## + Fold10.Rep07: alpha=0, lambda=0.5
## - Fold10.Rep07: alpha=0, lambda=0.5
## + Fold01.Rep08: alpha=0, lambda=0.5
## - Fold01.Rep08: alpha=0, lambda=0.5
## + Fold02.Rep08: alpha=0, lambda=0.5
## - Fold02.Rep08: alpha=0, lambda=0.5
## + Fold03.Rep08: alpha=0, lambda=0.5
## - Fold03.Rep08: alpha=0, lambda=0.5
## + Fold04.Rep08: alpha=0, lambda=0.5
## - Fold04.Rep08: alpha=0, lambda=0.5
## + Fold05.Rep08: alpha=0, lambda=0.5
## - Fold05.Rep08: alpha=0, lambda=0.5
## + Fold06.Rep08: alpha=0, lambda=0.5
## - Fold06.Rep08: alpha=0, lambda=0.5
## + Fold07.Rep08: alpha=0, lambda=0.5
## - Fold07.Rep08: alpha=0, lambda=0.5
## + Fold08.Rep08: alpha=0, lambda=0.5
## - Fold08.Rep08: alpha=0, lambda=0.5
## + Fold09.Rep08: alpha=0, lambda=0.5
## - Fold09.Rep08: alpha=0, lambda=0.5
## + Fold10.Rep08: alpha=0, lambda=0.5
## - Fold10.Rep08: alpha=0, lambda=0.5
## + Fold01.Rep09: alpha=0, lambda=0.5
## - Fold01.Rep09: alpha=0, lambda=0.5
## + Fold02.Rep09: alpha=0, lambda=0.5
## - Fold02.Rep09: alpha=0, lambda=0.5
## + Fold03.Rep09: alpha=0, lambda=0.5
## - Fold03.Rep09: alpha=0, lambda=0.5
## + Fold04.Rep09: alpha=0, lambda=0.5
## - Fold04.Rep09: alpha=0, lambda=0.5
## + Fold05.Rep09: alpha=0, lambda=0.5
## - Fold05.Rep09: alpha=0, lambda=0.5
## + Fold06.Rep09: alpha=0, lambda=0.5
## - Fold06.Rep09: alpha=0, lambda=0.5
## + Fold07.Rep09: alpha=0, lambda=0.5
## - Fold07.Rep09: alpha=0, lambda=0.5
## + Fold08.Rep09: alpha=0, lambda=0.5
## - Fold08.Rep09: alpha=0, lambda=0.5
## + Fold09.Rep09: alpha=0, lambda=0.5
## - Fold09.Rep09: alpha=0, lambda=0.5
## + Fold10.Rep09: alpha=0, lambda=0.5
## - Fold10.Rep09: alpha=0, lambda=0.5
## + Fold01.Rep10: alpha=0, lambda=0.5
## - Fold01.Rep10: alpha=0, lambda=0.5
## + Fold02.Rep10: alpha=0, lambda=0.5
## - Fold02.Rep10: alpha=0, lambda=0.5
## + Fold03.Rep10: alpha=0, lambda=0.5
## - Fold03.Rep10: alpha=0, lambda=0.5
## + Fold04.Rep10: alpha=0, lambda=0.5
## - Fold04.Rep10: alpha=0, lambda=0.5
## + Fold05.Rep10: alpha=0, lambda=0.5
## - Fold05.Rep10: alpha=0, lambda=0.5
## + Fold06.Rep10: alpha=0, lambda=0.5
## - Fold06.Rep10: alpha=0, lambda=0.5
## + Fold07.Rep10: alpha=0, lambda=0.5
## - Fold07.Rep10: alpha=0, lambda=0.5
## + Fold08.Rep10: alpha=0, lambda=0.5
## - Fold08.Rep10: alpha=0, lambda=0.5
## + Fold09.Rep10: alpha=0, lambda=0.5
## - Fold09.Rep10: alpha=0, lambda=0.5
## + Fold10.Rep10: alpha=0, lambda=0.5
## - Fold10.Rep10: alpha=0, lambda=0.5
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 0, lambda = 0.125 on full training set# print results
print(ridgeReg2)## glmnet
##
## 2938 samples
## 13 predictor
## 3 classes: 'bad', 'good', 'medium'
##
## No pre-processing
## Resampling: Cross-Validated (10 fold, repeated 10 times)
## Summary of sample sizes: 2645, 2645, 2644, 2644, 2645, 2643, ...
## Resampling results across tuning parameters:
##
## lambda Accuracy Kappa
## 0.000100 0.9952337 0.9878288
## 0.125075 0.9952337 0.9878288
## 0.250050 0.9948934 0.9869549
## 0.375025 0.9687897 0.9167402
## 0.500000 0.9454428 0.8490061
##
## Tuning parameter 'alpha' was held constant at a value of 0
## Accuracy was used to select the optimal model using the largest value.
## The final values used for the model were alpha = 0 and lambda = 0.125075.# plot results
plot(ridgeReg2)
plot(ridgeReg2$finalModel, xvar = 'lambda', lwd =1.4, label = TRUE)
plot(varImp(ridgeReg2, scale = FALSE))
plot(varImp(ridgeReg2, scale = TRUE))
PredictRidge2 <- predict(ridgeReg2, test)
confusionMatrix(PredictRidge2, test$quality_level)## Confusion Matrix and Statistics
##
## Reference
## Prediction bad good medium
## bad 73 0 0
## good 0 411 0
## medium 6 0 1470
##
## Overall Statistics
##
## Accuracy : 0.9969
## 95% CI : (0.9933, 0.9989)
## No Information Rate : 0.75
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.9921
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: bad Class: good Class: medium
## Sensitivity 0.92405 1.0000 1.0000
## Specificity 1.00000 1.0000 0.9878
## Pos Pred Value 1.00000 1.0000 0.9959
## Neg Pred Value 0.99682 1.0000 1.0000
## Prevalence 0.04031 0.2097 0.7500
## Detection Rate 0.03724 0.2097 0.7500
## Detection Prevalence 0.03724 0.2097 0.7531
## Balanced Accuracy 0.96203 1.0000 0.9939We have run the Ridge model on Quality Level variable. The accuracy is 99.69%
set.seed(123)
lassoReg1 <- train(quality~., train, method = 'glmnet',
tuneGrid = expand.grid(alpha = 1,
lambda = seq(0.0001, 0.5, length = 5)),
trControl = trainControl(method = "repeatedcv",
number = 10, repeats = 5,
verboseIter = TRUE))## + Fold01.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep1: alpha=1, lambda=0.5
## + Fold02.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep1: alpha=1, lambda=0.5
## + Fold03.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep1: alpha=1, lambda=0.5
## + Fold04.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep1: alpha=1, lambda=0.5
## + Fold05.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep1: alpha=1, lambda=0.5
## + Fold06.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep1: alpha=1, lambda=0.5
## + Fold07.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep1: alpha=1, lambda=0.5
## + Fold08.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep1: alpha=1, lambda=0.5
## + Fold09.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep1: alpha=1, lambda=0.5
## + Fold10.Rep1: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep1: alpha=1, lambda=0.5
## + Fold01.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep2: alpha=1, lambda=0.5
## + Fold02.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep2: alpha=1, lambda=0.5
## + Fold03.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep2: alpha=1, lambda=0.5
## + Fold04.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep2: alpha=1, lambda=0.5
## + Fold05.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep2: alpha=1, lambda=0.5
## + Fold06.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep2: alpha=1, lambda=0.5
## + Fold07.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep2: alpha=1, lambda=0.5
## + Fold08.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep2: alpha=1, lambda=0.5
## + Fold09.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep2: alpha=1, lambda=0.5
## + Fold10.Rep2: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep2: alpha=1, lambda=0.5
## + Fold01.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep3: alpha=1, lambda=0.5
## + Fold02.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep3: alpha=1, lambda=0.5
## + Fold03.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep3: alpha=1, lambda=0.5
## + Fold04.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep3: alpha=1, lambda=0.5
## + Fold05.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep3: alpha=1, lambda=0.5
## + Fold06.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep3: alpha=1, lambda=0.5
## + Fold07.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep3: alpha=1, lambda=0.5
## + Fold08.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep3: alpha=1, lambda=0.5
## + Fold09.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep3: alpha=1, lambda=0.5
## + Fold10.Rep3: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep3: alpha=1, lambda=0.5
## + Fold01.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep4: alpha=1, lambda=0.5
## + Fold02.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep4: alpha=1, lambda=0.5
## + Fold03.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep4: alpha=1, lambda=0.5
## + Fold04.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep4: alpha=1, lambda=0.5
## + Fold05.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep4: alpha=1, lambda=0.5
## + Fold06.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep4: alpha=1, lambda=0.5
## + Fold07.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep4: alpha=1, lambda=0.5
## + Fold08.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep4: alpha=1, lambda=0.5
## + Fold09.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep4: alpha=1, lambda=0.5
## + Fold10.Rep4: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep4: alpha=1, lambda=0.5
## + Fold01.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold01.Rep5: alpha=1, lambda=0.5
## + Fold02.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold02.Rep5: alpha=1, lambda=0.5
## + Fold03.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold03.Rep5: alpha=1, lambda=0.5
## + Fold04.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold04.Rep5: alpha=1, lambda=0.5
## + Fold05.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold05.Rep5: alpha=1, lambda=0.5
## + Fold06.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold06.Rep5: alpha=1, lambda=0.5
## + Fold07.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold07.Rep5: alpha=1, lambda=0.5
## + Fold08.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold08.Rep5: alpha=1, lambda=0.5
## + Fold09.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold09.Rep5: alpha=1, lambda=0.5
## + Fold10.Rep5: alpha=1, lambda=0.5## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground## - Fold10.Rep5: alpha=1, lambda=0.5
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 1, lambda = 1e-04 on full training set## Warning in lognet(x, is.sparse, ix, jx, y, weights, offset, alpha, nobs, :
## one multinomial or binomial class has fewer than 8 observations; dangerous
## ground# plot results
plot(lassoReg1)
plot(varImp(lassoReg1, scale = FALSE))
plot(varImp(lassoReg1, scale = TRUE))
PredictLasso1 <- predict(lassoReg1, test)
confusionMatrix(PredictLasso1, test$quality)## Confusion Matrix and Statistics
##
## Reference
## Prediction 3 4 5 6 7 8 9
## 3 0 2 0 0 0 0 0
## 4 6 71 0 0 0 0 0
## 5 0 0 310 160 0 0 0
## 6 0 0 278 721 0 0 0
## 7 0 0 0 0 337 73 1
## 8 0 0 0 0 0 0 0
## 9 0 0 0 1 0 0 0
##
## Overall Statistics
##
## Accuracy : 0.7342
## 95% CI : (0.714, 0.7536)
## No Information Rate : 0.45
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 0.598
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: 3 Class: 4 Class: 5 Class: 6 Class: 7 Class: 8
## Sensitivity 0.000000 0.97260 0.5272 0.8175 1.0000 0.00000
## Specificity 0.998976 0.99682 0.8834 0.7421 0.9544 1.00000
## Pos Pred Value 0.000000 0.92208 0.6596 0.7217 0.8200 NaN
## Neg Pred Value 0.996936 0.99894 0.8134 0.8325 1.0000 0.96276
## Prevalence 0.003061 0.03724 0.3000 0.4500 0.1719 0.03724
## Detection Rate 0.000000 0.03622 0.1582 0.3679 0.1719 0.00000
## Detection Prevalence 0.001020 0.03929 0.2398 0.5097 0.2097 0.00000
## Balanced Accuracy 0.499488 0.98471 0.7053 0.7798 0.9772 0.50000
## Class: 9
## Sensitivity 0.0000000
## Specificity 0.9994895
## Pos Pred Value 0.0000000
## Neg Pred Value 0.9994895
## Prevalence 0.0005102
## Detection Rate 0.0000000
## Detection Prevalence 0.0005102
## Balanced Accuracy 0.4997448We have run the Lasso model on Quality variable. The accuracy is 73.42%
lassoReg2 <- train(quality_level~., train, method = 'glmnet',
tuneGrid = expand.grid(alpha = 1,
lambda = seq(0.0001, 0.5, length = 5)),
trControl = trainControl(method = "repeatedcv",
number = 10, repeats = 5,
verboseIter = TRUE))## + Fold01.Rep1: alpha=1, lambda=0.5
## - Fold01.Rep1: alpha=1, lambda=0.5
## + Fold02.Rep1: alpha=1, lambda=0.5
## - Fold02.Rep1: alpha=1, lambda=0.5
## + Fold03.Rep1: alpha=1, lambda=0.5
## - Fold03.Rep1: alpha=1, lambda=0.5
## + Fold04.Rep1: alpha=1, lambda=0.5
## - Fold04.Rep1: alpha=1, lambda=0.5
## + Fold05.Rep1: alpha=1, lambda=0.5
## - Fold05.Rep1: alpha=1, lambda=0.5
## + Fold06.Rep1: alpha=1, lambda=0.5
## - Fold06.Rep1: alpha=1, lambda=0.5
## + Fold07.Rep1: alpha=1, lambda=0.5
## - Fold07.Rep1: alpha=1, lambda=0.5
## + Fold08.Rep1: alpha=1, lambda=0.5
## - Fold08.Rep1: alpha=1, lambda=0.5
## + Fold09.Rep1: alpha=1, lambda=0.5
## - Fold09.Rep1: alpha=1, lambda=0.5
## + Fold10.Rep1: alpha=1, lambda=0.5
## - Fold10.Rep1: alpha=1, lambda=0.5
## + Fold01.Rep2: alpha=1, lambda=0.5
## - Fold01.Rep2: alpha=1, lambda=0.5
## + Fold02.Rep2: alpha=1, lambda=0.5
## - Fold02.Rep2: alpha=1, lambda=0.5
## + Fold03.Rep2: alpha=1, lambda=0.5
## - Fold03.Rep2: alpha=1, lambda=0.5
## + Fold04.Rep2: alpha=1, lambda=0.5
## - Fold04.Rep2: alpha=1, lambda=0.5
## + Fold05.Rep2: alpha=1, lambda=0.5
## - Fold05.Rep2: alpha=1, lambda=0.5
## + Fold06.Rep2: alpha=1, lambda=0.5
## - Fold06.Rep2: alpha=1, lambda=0.5
## + Fold07.Rep2: alpha=1, lambda=0.5
## - Fold07.Rep2: alpha=1, lambda=0.5
## + Fold08.Rep2: alpha=1, lambda=0.5
## - Fold08.Rep2: alpha=1, lambda=0.5
## + Fold09.Rep2: alpha=1, lambda=0.5
## - Fold09.Rep2: alpha=1, lambda=0.5
## + Fold10.Rep2: alpha=1, lambda=0.5
## - Fold10.Rep2: alpha=1, lambda=0.5
## + Fold01.Rep3: alpha=1, lambda=0.5
## - Fold01.Rep3: alpha=1, lambda=0.5
## + Fold02.Rep3: alpha=1, lambda=0.5
## - Fold02.Rep3: alpha=1, lambda=0.5
## + Fold03.Rep3: alpha=1, lambda=0.5
## - Fold03.Rep3: alpha=1, lambda=0.5
## + Fold04.Rep3: alpha=1, lambda=0.5
## - Fold04.Rep3: alpha=1, lambda=0.5
## + Fold05.Rep3: alpha=1, lambda=0.5
## - Fold05.Rep3: alpha=1, lambda=0.5
## + Fold06.Rep3: alpha=1, lambda=0.5
## - Fold06.Rep3: alpha=1, lambda=0.5
## + Fold07.Rep3: alpha=1, lambda=0.5
## - Fold07.Rep3: alpha=1, lambda=0.5
## + Fold08.Rep3: alpha=1, lambda=0.5
## - Fold08.Rep3: alpha=1, lambda=0.5
## + Fold09.Rep3: alpha=1, lambda=0.5
## - Fold09.Rep3: alpha=1, lambda=0.5
## + Fold10.Rep3: alpha=1, lambda=0.5
## - Fold10.Rep3: alpha=1, lambda=0.5
## + Fold01.Rep4: alpha=1, lambda=0.5
## - Fold01.Rep4: alpha=1, lambda=0.5
## + Fold02.Rep4: alpha=1, lambda=0.5
## - Fold02.Rep4: alpha=1, lambda=0.5
## + Fold03.Rep4: alpha=1, lambda=0.5
## - Fold03.Rep4: alpha=1, lambda=0.5
## + Fold04.Rep4: alpha=1, lambda=0.5
## - Fold04.Rep4: alpha=1, lambda=0.5
## + Fold05.Rep4: alpha=1, lambda=0.5
## - Fold05.Rep4: alpha=1, lambda=0.5
## + Fold06.Rep4: alpha=1, lambda=0.5
## - Fold06.Rep4: alpha=1, lambda=0.5
## + Fold07.Rep4: alpha=1, lambda=0.5
## - Fold07.Rep4: alpha=1, lambda=0.5
## + Fold08.Rep4: alpha=1, lambda=0.5
## - Fold08.Rep4: alpha=1, lambda=0.5
## + Fold09.Rep4: alpha=1, lambda=0.5
## - Fold09.Rep4: alpha=1, lambda=0.5
## + Fold10.Rep4: alpha=1, lambda=0.5
## - Fold10.Rep4: alpha=1, lambda=0.5
## + Fold01.Rep5: alpha=1, lambda=0.5
## - Fold01.Rep5: alpha=1, lambda=0.5
## + Fold02.Rep5: alpha=1, lambda=0.5
## - Fold02.Rep5: alpha=1, lambda=0.5
## + Fold03.Rep5: alpha=1, lambda=0.5
## - Fold03.Rep5: alpha=1, lambda=0.5
## + Fold04.Rep5: alpha=1, lambda=0.5
## - Fold04.Rep5: alpha=1, lambda=0.5
## + Fold05.Rep5: alpha=1, lambda=0.5
## - Fold05.Rep5: alpha=1, lambda=0.5
## + Fold06.Rep5: alpha=1, lambda=0.5
## - Fold06.Rep5: alpha=1, lambda=0.5
## + Fold07.Rep5: alpha=1, lambda=0.5
## - Fold07.Rep5: alpha=1, lambda=0.5
## + Fold08.Rep5: alpha=1, lambda=0.5
## - Fold08.Rep5: alpha=1, lambda=0.5
## + Fold09.Rep5: alpha=1, lambda=0.5
## - Fold09.Rep5: alpha=1, lambda=0.5
## + Fold10.Rep5: alpha=1, lambda=0.5
## - Fold10.Rep5: alpha=1, lambda=0.5
## Aggregating results
## Selecting tuning parameters
## Fitting alpha = 1, lambda = 1e-04 on full training set# plot results
plot(lassoReg2)
plot(lassoReg2$finalModel, xvar = 'lambda', lwd =1.4, label=TRUE)
plot(varImp(lassoReg2, scale = FALSE))
plot(varImp(lassoReg2, scale = TRUE))
PredictLasso2 <- predict(lassoReg2, test)
confusionMatrix(PredictLasso2, test$quality_level)## Confusion Matrix and Statistics
##
## Reference
## Prediction bad good medium
## bad 79 0 0
## good 0 411 0
## medium 0 0 1470
##
## Overall Statistics
##
## Accuracy : 1
## 95% CI : (0.9981, 1)
## No Information Rate : 0.75
## P-Value [Acc > NIR] : < 2.2e-16
##
## Kappa : 1
## Mcnemar's Test P-Value : NA
##
## Statistics by Class:
##
## Class: bad Class: good Class: medium
## Sensitivity 1.00000 1.0000 1.00
## Specificity 1.00000 1.0000 1.00
## Pos Pred Value 1.00000 1.0000 1.00
## Neg Pred Value 1.00000 1.0000 1.00
## Prevalence 0.04031 0.2097 0.75
## Detection Rate 0.04031 0.2097 0.75
## Detection Prevalence 0.04031 0.2097 0.75
## Balanced Accuracy 1.00000 1.0000 1.00We have run the Lasso model on Quality Level variable. The accuracy is 100%